Discov Oncol. 2026 Jul 4. doi: 10.1007/s12672-026-05522-y. Online ahead of print.
ABSTRACT
BACKGROUND: Colorectal cancer (CRC) is a prevalent malignant tumor with increasing incidence and mortality rates worldwide. Exosomes are secretory vesicles generated by the endosomal system within cells. Previous studies have reported that exosome-related genes (ERGs) are associated with the progression of malignancies. This study investigates the role of ERGs in CRC, evaluates their impact on CRC prognosis, and explores inter-individual differences among CRC patients in different risk groups.
METHODS: Weighted Gene Co-expression Network Analysis (WGCNA) algorithm was employed to identify ERGs associated with CRC. Subsequently, various bioinformatics approaches, including enrichment analysis, consensus clustering, and survival analysis, were utilized to investigate the role of ERGs in individual CRC patients. Furthermore, exosome-related signature genes were refined using the Random Forest (RF) and Least Absolute Shrinkage and Selection Operator (LASSO) algorithms based on colon cancer samples from The Cancer Genome Atlas (TCGA) database. An ERG-related gene signature was then constructed to calculate the ERG-associated risk score for each patient, which was validated using data from the Gene Expression Omnibus (GEO) database. Based on the risk scores, we assessed the responsiveness of different CRC individuals to immunotherapy and chemotherapy. Finally, single-cell analysis provided deeper insights into the relationship between ERGs and CRC, and a nomogram was established to enhance their clinical utility.
RESULTS: This study identified two ERG-related subtypes that exhibited significant differences in prognosis, enriched pathways, clinicopathological features, and immune characteristics. Moreover, CRC individuals with high ERG-related risk scores were associated with poor responsiveness to immunotherapy and increased sensitivity to various chemotherapeutic agents. Single-cell analysis revealed that ERGs were highly expressed in monocytes. The model developed in this study demonstrated strong predictive accuracy for assessing ERG-related risk in CRC patients.
CONCLUSIONS: This study identified two hub genes, CUL4A and UCHL1, highlighting the diagnostic and prognostic significance of ERGs in CRC and offering new insights into CRC treatment. Additionally, the ERG signature plays a crucial role in predicting individualized prognosis and facilitating the development of novel therapeutic strategies for CRC patients. Nevertheless, further studies are necessary to bring statistically derived risk models into clinically applicable assays.
PMID:42400715 | DOI:10.1007/s12672-026-05522-y
